Abstract. The World Wide Web is an enormous and a growing source of information presented in a human friendly language called Html. Unfortunately, querying and accessing this information by software agents is not an easy task, so web information extractors are used. Currently, there is a variety of algorithms to build web information extractors, but none of them is universally applicable. There is not a common software framework to develop them. This has resulted in proposals that range in complexity, precision and recall, but having diverging interfaces, which makes it difficult to reuse or integrate them. As a result, few side-by-side comparisons are available, but none of them is complete. We argue that the key is the absence of a unifying framework in which researchers can develop their proposals so that they can be assessed properly. Devising and implementing such a framework would be an ultimate tool to help reduce costs at integrating web information into automatic business processes. In this paper we report on our first version of this framework for information extractors. Key words: Information extraction, Enterprise Information Integration. 1